Sensitivity and specificity for the first and last segments of spliced 5′ UTRs (E5′FIRST and E5′LAST) as well as unspliced 5′ UTRs (E5′SNGL), evaluated by comparison to aligned sequences from DBTSS.
Match | Overlap | ||||||
|---|---|---|---|---|---|---|---|
| Feature | 5′ UTR Type | Sn (%) | Sp (%) | Sn (%) | Sp (%) | ||
| E5′FIRST | Spliced | 25 | 25 | 43 | 43 | ||
| E5′LAST | Spliced | 42 | 43 | 54 | 54 | ||
| E5′SNGL | Unspliced | 66 | 68 | 67 | 70 | ||
[i] The match (both boundaries are true positives) and overlap (the predicted feature overlaps the annotation by at least one base) values are displayed. Matches are a special case of overlap and hence they contribute to the overlap total. All specificity values are scaled assuming that there are 25,000 genes in the human genome and that the annotation set is a random sample of them. Rounded to the nearest percentage, sensitivity and specificity values are equal for match and overlap for both features.